Cold Regions Science and Technology 120 (2015) 127–137

Contents lists available at ScienceDirect

Cold Regions Science and Technology

j ourna l homepage: www.e lsev ie r .com/ locate /co ld reg ions

Effects of freeze–thaw cycles on a fiber reinforced fine grained soil inrelation to geotechnical parameters

Mahya Roustaei ⁎, Abolfazl Eslami, Mahmoud GhazaviCivil Engineering Department, Islamic Azad University, Qazvin Branch, IranDepartment of Civil and Environmental Engineering, Amirkabir University of Technology (AUT), Tehran, IranFaculty of Civil Engineering, K. N. Toosi University of Technology, Tehran, Iran

⁎ Corresponding author. Tel.: +98 28 33665275; fax: +E-mail address: Mahya_roustaei@qiau.ac.ir (M. Rousta

http://dx.doi.org/10.1016/j.coldregions.2015.09.0110165-232X/© 2015 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 20 August 2014Received in revised form 6 September 2015Accepted 11 September 2015Available online 30 September 2015

Keywords:Freeze–thaw cyclingFine grained soilPolypropylene fibersUU triaxial testSEM images

Freeze–thaw cycling is a weathering process which occurs in cold climates during winter and spring. At temper-atures just below 0 °C, ice lenseswhich tend to form in free spaces between soil aggregates, force them apart andend up the alteration of characteristic structures in micro and macro scales. In most of the previous studies,changes in physical and chemical properties of soils were investigated. This study was conducted to manifestthe effect of using polypropylene fibers in a fine grained soil during freeze–thaw cycles. A clayey soil, reinforcedwith 0.5, 1 and 1.5 percentages of polypropylene fibers, was compacted in the laboratory and exposed to a max-imum of 9 closed-system freeze–thaw cycles. It has been found that for the investigated soil, unconsolidated un-drained triaxial compressive strength of unreinforced soil decreases with increasing the number of freeze–thawcycles, whereas reinforced sample shows better performance and the strength reduction amount decreases from43% to 32% by reinforcing the soil samples. This effect is caused by acting polypropylenefibers as tensile elementsbetween the soil particles as it is demonstratedwith scanning electronmicroscope (SEM). In addition reinforcingcan also reduces the effect of freeze–thaw cycles on the changes of cohesion of the soil.

© 2015 Elsevier B.V. All rights reserved.

1. Introduction

In cold climates, soils, especially on shallow depths, are exposed tofreeze–thaw cycles every year. These cycles considerably change the en-gineering properties of soils. During soil freezing, ice lenses which tendto form in free spaces between soil aggregates, force them apart and endup the alteration of characteristic structures in micro and macro scales.

The conditions of ice segregation in frozen soils mainly depend onfreezing factors such as temperature, soil types and available freewater.

Soil behavior due to freeze–thaw cycles was investigated in numer-ous studies as in the permafrost regions these cycles reduced the suffi-ciency of soil structures. In Canada it has been found that theembankment constructed on soil which has never experienced freeze–thaw cycles, was damaged in just one year due to the loss of bearing ca-pacity (Leroueil et al., 1991). Therefore, newly constructedhighway em-bankments that are left unpaved for a few years may be subjected topossible damages by freeze–thaw cycles (Eigenbrod, 1996).

Qi et al. (2006) reviewed the latest effortswhichwere done to inves-tigate the influence of freeze–thaw cycles on soil properties. They sum-marized these influences in two parts: physical properties such asdensity and hydraulic permeability and mechanical properties such asultimate strength, strain–stress behavior and resilient modulus. As

98 28 33675784.ei).

mentioned in this research, loose soils tend to be densified and densesoils become looser after freeze–thaw cycles and both loose and densesoils may attain the same void ratio after a number of cycles (Konrad,1989). By increasing the large pores that are left after the thaw of icecrystals, permeabilitywill increase (Chamberlain et al., 1990). These cy-cles reduce the ultimate strength of soils. All over-consolidated soils ex-hibit a peak on the triaxial stress–strain curve that is reduced or mayeven disappear (Graham and Au, 1985). Resilient modulus is one ofthe most important factors in pavement designs that will be reducedsignificantly by even a small number of freeze–thaw cycles (Simonsenand Isacsson, 2001). In addition, these cycles decrease the undrainedshear strength considerably which is an important factor in engineeringproperties of fine-grained soils (Graham and Au, 1985).

It is worth mentioning that the changes of soil microscopic proper-ties during freeze–thaw cycles result in the changes of engineeringcharacteristics of soils. These microstructural changes have been inves-tigated through SEM (scanning electron microscope) and it was foundthat a very significant increase of the permeability of clayey soil was ob-served after freezing and thawing (Hohmann-Porebska, 2002) and alsothe soil becomes looser as the equivalent diameter decreases (Cui et al.,2014).

In addition to static mechanical parameters of soil, the dynamiccharacteristic changes of soil have been recently considered duringfreeze–thaw cycles.Wang et al. (2015) found that the dynamicmodulusof a silty soil greatly decreases, whereas the damping ratio increases

Table 1Properties of the soil.

Gs (specific gravity) 2.657Plastic limit 36%Liquid limit 20%Plastic index 16%

128 M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

with additional freeze–thaw cycles; and changes level off after the sixthcycle. The dynamic properties after 6–7 freeze–thaw cycles are sug-gested for use in designing and calculating indexes. These results arein agreement with the findings of Cui et al. (2014) who performed dy-namic tests on a silty clay.

All above-mentioned research works deal with unreinforced soil.Alongwithmany applications for soil improvement, there come severalwidely varied methods. Taking the influence of freeze–thaw cycles onsoils into consideration, just a few researchers have contemporarily fo-cused on using additives which can control the effects of these cycles.

Yarbesi et al. (2007) stabilized two granular soils by silica fume–lime, fly ash–lime, and red mud–cement. Their experimental resultsshowed that stabilized samples with the mentioned additive mixtureshave high freezing–thawing durability as compared to unstabilizedsamples. These additive mixtures which have also improved the dy-namic behavior of the soil samples can be successfully used as an addi-tive material to enhance the freeze–thaw durability of granular soils forroad constructions and earthwork applications.

Kalkan (2009) used a fine grained soil stabilized by adding silicafumewhichwas generated during siliconmetal production. The test re-sults showed that the stabilized fine-grained soil exhibits high resis-tance to the freezing and thawing effects as compared to natural fine-grained soil samples. The silica fume decreases the effects of freeze–thaw cycles on unconfined compressive strength and permeability.

Hazirbaba and Gullu (2010) performed CBR tests to investigate theinfluence of freeze–thaw conditions and also no freeze–thaw conditionson fine-grained soil samples which were treated with the inclusion ofgeofiber and synthetic fluid in soaked and unsoaked conditions. The re-sults indicated that the addition of geofiber togetherwith syntheticfluidis generally successful in providing resistance against freeze–thawweakening. However, the addition of synthetic fluid alone is not very ef-fective against the detrimental impact of freeze–thaw cycles. The resultsfrom soaked samples subjected to a freeze–thaw cycle show poor CBRperformance for treatments involving syntheticfluid,while samples im-proved with geofibers alone generally offer better performance. Liu etal. (2010) conducted dynamic triaxial tests on cement and lime-modi-fied soils with different blend ratios in freeze–thaw cycles. The resultsshowed that after repeated freeze–thaw cycles, the modified soils ex-hibit better performance than before modification, the cement-modi-fied clay is superior to the lime modified clay, and all soil mechanicalproperties are visibly improved.

Zaimoglu (2010) investigated the effect of randomly distributedpolypropylene fibers on strength and durability behavior of a fine-grained soil subjected to freezing–thawing cycles. The content of poly-propylene fiber varied between 0.25% and 2% by dry weight of soil inthe tests. It was observed that the mass loss in reinforced soils is almost50% lower than that in unreinforced soil. It was also found that the un-confined compressive strength of specimens subjected to freezing–thawing cycles generally increases with increasing fiber content. In ad-dition, the results indicated that the initial stiffness of the stress–straincurves is not affected significantly by the fiber reinforcement in the un-confined compression tests.

Ghazavi and Roustaie (2010) reinforced a caolinite clay with steeland polypropylene fibers and exposed the soil samples to a maximumof 10 closed-system freeze–thaw cycles. They found that increasingthe number of freeze–thaw cycles results in the decrease of unconfinedcompressive strength of clay samples by 20%–25%. Moreover, the inclu-sion of fiber in clay samples increases the unconfined compressivestrength of soil and decreases the frost heave. Furthermore, the resultsof the study indicated that the addition of 3% polypropylene fibers re-sults in the increase of unconfined compressive strength of the soil be-fore and after applying freeze–thaw cycles by 60% to 160% anddecrease of frost heave by 70%.

The unconfined compression tests have been conducted on clay–polypropylene mixtures after freeze–thaw cycles in previously men-tioned researches. As the confining pressure is a factor which causes

soil particles to move, rearrange, consolidate, and recover soil strengthin a sense, it can be an important factor in determination of soil strengthafter freeze–thaw cycles (Wang et al, 2007).

In 2014, the stabilizationmethod of soil, affected by freeze–thaw cy-cles, attracted increasing attention. Unconfined compressive strength(UCS) of gypsum soil samples (gypsum content of 5%, 10% and 25%) de-creases greatly, and samples loose substantially all of their strengthfrom the 5th cycle but the lime treated samples without gypsum revealbetter durability to freeze–thaw cycles (Aldaood et al, 2014).

The UCS increases with increased bassanite and coal ash contents inthe soil.With respect to freezing and thawing durability, the first or sec-ond cycle of freeze–thaw,markedly decreases the unconfined compres-sive strength of both treated and untreated cement stabilized soils, butfurther cycles have little additional influence (Shibi and Kamei, 2014).Finally Güllü and Khudir (2014) showed that although the potential ef-fective rates of the stabilizers are found to be 0.75% jute fiber, 0.25% steelfiber and 4% lime for soil stabilization but as the freeze–thaw cycles in-crease, the UCS values decrease at the treatments, except for the addi-tions of jute fiber alone. The jute fiber inclusions relatively decreasethe brittleness index toward zero at all freeze–thaw cycles.

Having these results in mind, the present study investigates the ef-fect of confining pressure on the behavior of a fine grained reinforcedsoil during freeze–thaw cycles. To this end, the strength of the soil rein-forced with 0.5%, 1% and 1.5% of polypropylene fibers was evaluatedthrough triaxial test with three different confining pressures (30, 60,and 90 kPa) before and after 0, 1, 3, 6, and 9 freeze–thaw cycles.

2. Materials

In this study, a fine grained soil, classified as CL in the Unified SoilClassification System, underwent laboratory tests. Noticeably, the ef-fects of freeze–thaw cycles are more considerable in fine-grain soils incomparison with sand or gravel (Qi et al., 2006).

The soil properties are presented in Table 1 and its grain size distri-bution is shown in Fig. 1. Standard Proctor Compaction tests were per-formed on the soil, and a maximum dry mass density ofapproximately 1.78 g/cm3 at optimum moisture content (OMC) of ap-proximately 17.4% was obtained. The specimens are reinforced using0.5%, 1% and 1.5% of polypropylene fiber contents of weight of dry soil.The properties of the fibers are presented in Table 2.

3. Experimental procedure

This investigation aims at studying the effects of application of poly-propylene fibers on the strength changes of highly compressible finegrained soil compacted at maximum dry density with the optimummoisture content and subjected to 0–9 freeze–thaw cycles.

In order to find out the amount of changes of the soil in these cycles,fourmain following steps should be taken for each sample. Some verifi-cation tests are also carried out in order to examine the repeatability ofthe experiment results.

3.1. Sample preparation

All cylindrical samples with 50 mm diameter and 100 mm heightwere prepared with the maximum dry unit weight and optimumwater content. To prepare samples, firstly, the necessary OMC wasdetermined and mixed with the soil. The soil and the fiber amounts

Fig. 1. Grain size distribution of the fine grained soil.

129M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

were divided into 4 parts and compacted in themold. After the removalof each sample from the mold, the sample is immediately covered witha plastic layer which helps it against water evaporation.

3.2. Freeze–thaw cycles

To prepare the samples for the closed system freezing and thawingcycles, specimens were placed in a digital refrigerator at −20 °C for6 h and then at +20 °C for thawing phase for 6 h. These temperatureshad been previously used in some researches (Qi et al., 2006). Sixhours is a proportional period after which the alteration of specimens'height would become constant. This means that the height increase infreeze phase and the height decrease in thaw phase stop. The cycleswere continued up to 9 cycles. This number of cycles was chosen sincemost soil strength reduction would occur in primary cycles and after5–10 cycles a new equilibrium condition would become predominanton samples (Ghazavi and Roustaie, 2010).

3.3. Strength testing

After freeze–thaw cycles, strength parameters of the soil were mea-sured in unconsolidated undrained (UU) triaxial compression tests. Inaccordance with ASTM D2850-03a, throughout the testing programthe strain rate kept constant at 1 mm per minute.

As the freeze–thaw cycles often occur in surface parts of groundwhich is specially connected with road pavement and is subjectedwith rapid loading from vehicles, the fine-grained soil with low perme-ability cannot be drained and consolidated consequently. Therefore, theUU test is suitable for investigating the behavior of thawing soil subject-ed to rapid and repeated traffic load at shallow surface.

To simulate these conditions at shallowdepths of soil, three differentconfining pressures of 30, 60, and 90 kPa have been selected for triaxialtests.

3.4. SEM images

A scanning electron microscope (SEM) is a type of electron micro-scope that produces images of a sample by scanning it with a focusedbeam of electrons. One of the most surprising aspects of SEM is the

Table 2Properties of polypropylene fibers.

Fiber Polypropylene

Length (mm) 12Diameter (mm) 0.1Unit weight (kg/m3) 900

apparent ease with which images of three-dimensional objects can beinterpreted by any observer with no prior knowledge of the instrument.This is somewhat amazing in view of the unusualway inwhich image isformed, which seems to differ greatly from normal human experiencewith images formed by light and viewed by eyes. In this study the alter-ations of reinforced and unreinforced fine grained soil samples were di-rectly investigated before and after the cycles by analyzing the SEMimages which were taken from the soil samples.

For the purpose of investigating the interfacial interactions betweenthe fiber surface and soil matrix, several related SEM images are givenfrom the pure and reinforced soil specimen before and after thefreeze–thaw cycles. Discussing the details of these pictures in the fol-lowing sections results in finding the micro-structural changes of thesoil and interactions between the soil particles and fibers during thecycles.

4. Results

In order to find out the detailed changes of mechanical behavior ofpure and reinforced soil influenced by freeze–thaw cycles, triaxial com-pression tests have been conducted on unfrozen and thawed soil underthree different confining pressures after the soil was subjected to 1, 3, 6,and 9 cycles of freeze–thaw.

The variations of stress–strain response of samples are named NC-UnRe for unreinforced samples under N freeze–thaw cycles and NC-XPRe for reinforced samples with X% of polypropylene fibers subjectedto N freeze–thaw cycles.

4.1. Effect of using polypropylene fibers on mechanical behavior of the soilbefore freeze–thaw cycles

Strength of the soil is measured using the UU triaxial test accordingto the methodology described in ASTM D2850-03a and the changes ofsoil mechanical features after freeze–thaw cycles were investigated. Be-fore any cycles, it is recommended to measure the changes in the finegrained soil properties due to reinforcing with fibers. Adding polypro-pylene fibers to soil samples increases the strength considerably as itis shown in Fig. 2. This figure indicates the stress–strain curves of rein-forced specimen with 0, 0.5, 1 and 1.5% of fibers in the confining pres-sures of 30, 60 and 90 kPa.

By applying three confining pressures in UU triaxial test and consid-ering the peak of stress–strain curves or the maximum stress till 20%axial strain as failure strength, the effect of using polypropylene fibersin increasing the strength of the fine grained soil is clarified in Fig. 3.

It is obvious that the peak stresswas generally gone up by increasingfiber content. In comparison with the unreinforced sample, the com-pression strength of the reinforced sample at 1.5% polypropylene fiber

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

a

0C-0.5P Re0C-Un Re0C-1P Re0C-1.5P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

b

0C-0.5P Re0C-Un Re0C-1P Re0C-1.5P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

c

0C-Un Re0C-0.5P Re0C-1P Re0C-1.5P Re

Fig. 2. Stress–strain variation of unreinforced and reinforced samples before freeze–thawcycles under three different confining pressures: a) 30 kPa, b) 60 kPa, and c) 90 kPa.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0.03 0.06 0.09

Ult

imat

e st

ren

gth

(M

Pa)

Pressure (MPa)

UnRe0.5P-Re1P-Re1.5P-Re

Fig. 3. Variation of ultimate strength of unreinforced and reinforced samples versus con-fining pressures.

130 M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

content increased from 0.73 to 1.18MPa. It can be concluded that inclu-sion of 0.5%–1.5% polypropylene fibers results in increasing the failurestrength to about 28%–65%.

It can also be seen that thefiber-reinforced soil exhibitsmore ductilebehavior than the unreinforced soil. Furthermore, in very small strains,the initial stiffness of the soil appears not to be affected by the additionof fiber. Similar results were also obtained for high plasticity silt(Zaimoglu, 2010).

Cohesion is one of the most important characteristics of clayey soilswhich considerably effects the strength of cohesive soils as clay. The in-fluence of adding fibers on the cohesion of CL samples is illustrated inFig. 4. As seen, the soil cohesion increases about 67%–100% by adding fi-bers, but by increasing the fiber content this increment is reduced. Thecohesion growth is a result of strength increase in soil samples due tofiber addition and plasticity. However, polypropylene fibers, becauseof their polymeric material, do not have significant cohesion with soilparticles. So the cohesion reduction can be visible by increasing thefiber content in the soil.

4.2. Effect of freeze–thaw cycles on the reinforced and unreinforced soil

Strength features of reinforced and unreinforced soil weremeasuredin UU triaxial compression tests after 0, 1, 3, 6, and 9 cycles. The resultsare described in the following sections.

4.2.1. Stress–strain curvesSeparation of soil aggregates, which is caused by ice lenses made up

of soil pure water at temperatures just below 0 °C, disrupts theinterlocking of soil grains and changes the mechanical properties ofsoil. Figs. 5–8 show the test results for reinforced and unreinforced sam-ples subjected to 0, 1, 3, 6, and 9 freeze–thaw cycles in three differentconfining pressures. As seen, in all reinforced and unreinforced samples,by increasing the number of freeze–thaw cycles, the soil strength de-creases. However, the strength decrease is more visible in unreinforcedsamples than in reinforced ones. The stress–strain variation of thawedsoil tends to vary from strain-hardening type to strain-softening type.The strength reduction for reinforced and unreinforced soil subjectedto freeze–thaw cycles was also observed in previous studies (Ghazaviand Roustaie, 2010, 2013; Graham and Au, 1985).

4.2.2. Failure strengthBy considering the peak of stress–strain curves as failure strength,

Fig. 9 shows the triaxial strength ratio of reinforced and unreinforcedsamples versus the number of freeze–thaw cycles for various confiningpressures. The ratio is defined as the strength of a reinforced or unrein-forced sample at a given cycle divided by that of the same samplewhichis not subjected to freeze–thaw cycles. The strength values are denotedby SN and S0, respectively. (See Figs. 5– 8.)

It is obvious from this figure that by increasing the number offreeze–thaw cycles, the strength of both reinforced and unreinforced

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Co

hes

ion

(M

Pa)

UnRe Re-0.5P Re-1P Re-1.5P

Fig. 4. Variation of cohesion of clay after fiber addition.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

a

0C-Un Re 1C-Un Re

3C-Un Re 6C-Un Re

9C-Un Re

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 5 10 15 20

Dev

iato

ri S

tres

s (M

Pa)

Strain (%)

b

0C-Un Re 1C-Un Re

3C-Un Re 6C-Un Re

9C-Un Re

0

0.2

0.4

0.6

0.8

1

1.2

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

c

0C-Un Re 1C-Un Re3C-Un Re 6C-Un Re9C-Un Re

Fig. 5. Stress–strain variation of unreinforced samples after freeze–thaw cycles underthree different confining pressures: a) 30 kPa, b) 60 kPa, and c) 90 kPa.

0

0.2

0.4

0.6

0.8

1

1.2

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

a

0C-0.5P Re 1C-0.5P Re3C-0.5P Re 6C-0.5P Re9C-0.5P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20 25

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

b

0C-0.5P Re 1C-0.5P Re3C-0.5P Re 6C-0.5P Re9C-0.5P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

c

0C-0.5P Re 1C-0.5P Re3C-0.5P Re 6C-0.5P Re9C-0.5P Re

Fig. 6. Stress–strain variation of reinforced samples with 0.5% fibers after freeze–thaw cy-cles under three different confining pressures: a) 30 kPa, b) 60 kPa, and c) 90 kPa.

131M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

samples decreases. In addition, it is clear that by increasing the confin-ing pressure, the strength reduction decreases. Therefore, freeze–thawcycles are more destructive on surface of the groundwhich is in contactwith structure foundations or road pavement.

Moreover, the soil particle rearrangementwhich is caused by higherconfining pressure will close the cracks and fissures produced byfreeze–thaw process and enhance the soil strength (Wang et al.,2007). So the effect of freeze–thaw cycles is more considerable in lowconfining pressures than high ones.

In addition, although the effect of freeze–thaw cycleswas investigat-ed in previous studies on reinforced soil with polypropylene fibers bythe authors, no effect was found in the absence of confining pressureand unconfined compression strength of reinforced soil reduced duringthe cycles. This phenomenon is a result of scanty cohesion between

fibers and soil particles which can be reduced by using confining pres-sure (Ghazavi and Roustaie, 2010).

According to these curves, using propylene fibers in the clayey soilaffects the strength reduction caused by freeze–thaw cycles. The com-pression strength of unreinforced samples decreases by 14%–43% dueto the application of 9 freeze–thaw cycles while this reduction isabout 1%–32% for propylene-reinforced samples.

4.2.3. Resilient modulusResilient modulus is a fundamental material property used to char-

acterize unbound pavement materials. It is a measure of material stiff-ness and provides a mean to analyze stiffness of materials underdifferent circumstances, such as moisture content, density, and stresslevel. It is also a required input parameter for mechanistic–empiricalpavement design method.

0

0.2

0.4

0.6

0.8

1

1.2

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

a

0C-1P Re 1C-1P Re3C-1P Re 6C-1P Re9C-1P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

b

0C-1P Re 1C-1P Re

3C-1P Re 6C-1P Re

9C-1P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

c

0C-1P Re 1C-1P Re

3C-1P Re 6C-1P Re

9C-1P Re

Fig. 7. Stress–strain variation of reinforced sampleswith 1%fibers after freeze–thaw cyclesunder three different confining pressures: a) 30 kPa, b) 60 kPa, and c) 90 kPa.

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

a

0C-1.5P Re 1C-1.5P Re

3C-1.5P Re 6C-1.5P Re

9C-1.5P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20

Dev

iato

r S

tres

s (M

Pa)

Strain (%)

b

0C-1.5P Re 1C-1.5P Re

3C-1.5P Re 6C-1.5P Re

9C-1.5P Re

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

0 5 10 15 20

Dev

iato

ri S

tres

s (M

Pa)

Strain (%)

c

0C-1.5P Re 1C-1.5P Re3C-1.5P Re 6C-1.5P Re9C-1.5P Re

Fig. 8. stress–strain variation of reinforced samples with 1.5% fibers after freeze–thaw cy-cles under three different confining pressures: a) 30 kPa, b) 60 kPa, and c) 90 kPa.

132 M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

The resilient modulus is defined as a ratio of the deviator stress in-crement at 1% axial strain to the axial strain increment, which can beexpressed as:

E ¼ ΔσΔε

¼ σ1:0%−σ0

ε1:0%−ε0ð1Þ

whereΔσ is the increment of deviator stress,Δε is the increment of axialstrain; σ1.0% is the deviator stress corresponding to the axial strain of1.0% (ε1.0%); and σ0 and ε0 are the initial stress and strain, respectively(Wang et al., 2007).

Lee et al. (1995) investigated the resilient properties of cohesivesoils and found that cohesive soils with resilient modulus lower than55 kPa would exhibit negligible freeze–thaw effects. In contrast, soilswith resilient modulus higher than 103 kPa would exhibit a decreaseof over 50% in this parameter due to freeze–thaw cycles.

Wang et al. (2007) also reported that the magnitude of the resilientmodulus decreases by 18%–27% of unfrozen soil depending on the con-fining pressure in triaxial compression tests.

The resilient modulus ratio of unreinforced and reinforced samplessubjected to each freeze–thaw cycle can be calculated by Eq. (1), as itis shown in Fig. 10. The ratio is defined as the resilient modulus of a re-inforced or unreinforced sample at a given cycle divided by that of thesame sample which is not subjected to freeze–thaw cycles. The resilientmodulus values are denoted by EN and E0, respectively.

The figure shows that by increasing the number of cycles, the resil-ient modulus of both reinforced and unreinforced samples decreasesand using the polypropylene fibers in the soil does not have significanteffect in controlling the reduction of this parameter. Fig. 10 also indi-cates that the decreased magnitude of resilient modulus is more obvi-ous in first cycles but when the number of freeze–thaw cycles exceedsseven, the resilient modulus will reach a certain value and remains

0

0.2

0.4

0.6

0.8

1

1.2

0 3 6 9

SN

/ S0

Cycle

UnRe-30Kpa0.5P-Re-30Kpa1P-Re-30Kpa1.5P-Re-30Kpa

0

0.2

0.4

0.6

0.8

1

1.2

0 3 6 9

SN

/S0

Cycle

UnRe-60Kpa

0.5P-Re-60Kpa

1P-Re-60Kpa

1.5P-Re-60Kpa

0

0.2

0.4

0.6

0.8

1

1.2

0 3 6 9

SN

/S0

Cycle

UnRe-90Kpa

0.5P-Re-90Kpa

1P-Re-90Kpa

1.5P-Re-90Kpa

Fig. 9. Variation of strength ratio of unreinforced and reinforced samples versus freeze–thaw cycles under confining pressures of 30, 60, and 90 kPa.

21

0

0.2

0.4

0.6

0.8

1

1.2

0 3 6 9

EN

/E0

Number of Cycle

Un Re-30 kPa0.5P-Re-30 kPa1P-Re-30 kPa1.5P-Re-30 kPa

0

0.2

0.4

0.6

0.8

1

1.2

1.4

EN

/E0

Number of Cycle

Un Re-60 kPa0.5P-Re-60 kPa1P-Re-60 kPa1.5P-Re-60 kPa

0

0.2

0.4

0.6

0.8

1

1.2

EN

/E0

Number of Cycle

Un Re-90 kPa0.5P-Re-90 kPa1P-Re-90 kPa1.5P-Re-90 kPa

0 3 6 9

0 3 6 9

Fig. 10. Variation of resilient modulus of unreinforced and reinforced samples versusfreeze–thaw cycles under confining pressures of 30, 60, and 90 kPa.

133M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

constant upon applying further freeze–thaw cycles. This phenomenonshows a new equilibrium condition which becomes predominant onthe fine grained soil after 6–7 cycles as most of the changes occur at1st to 7th cycles (Ghazavi and Roustaie, 2010; Qi et al., 2006).

0

0.4

0.8

1.2

1.6

2

2.4

0 3 6 9

CN

/C0

Number of Cycle

Un Re

0.5P-Re

1P-Re

1.5P-Re

Fig. 11. Variation of soil cohesion of unreinforced and reinforced samples versus freeze–thaw cycles.

4.2.4. CohesionCohesion is an important factor for evaluating the shear strength of

fine-grained soils. This parameter reflects the synthesis action of allkinds of physical–chemical forces between particles, such as Coulombforce, Van der Waals forces and the ions of adjacent particles, bondingaction and so on. The magnitude of cohesion is frequently influencedby the space between particles and bonding force caused by bondingmaterial (Wang et al., 2007).

In order to investigate the effect of freeze–thaw action on shearstrength of the CL soil which was chosen in this study, the impact ofthese cycles on the cohesive force will be observed.

The influence made by the number of freeze–thaw cycles on the co-hesion ratio of unreinforced and reinforced samples is illustrated inFig. 11. The ratio is defined as the cohesion of a reinforced or unrein-forced sample at a given cycle divided by that of the same samplewhich is not subjected to freeze–thaw cycles. The cohesion values aredenoted by CN and C0, respectively.

0

0.4

0.8

1.2

1.6

2

2.4

2.8

0 3 6 9

φ(N

)/φ(

0)

Number of Cycle

Un Re

0.5P-Re

1P-Re

1.5P-Re

Fig. 12. Variation of friction angle of unreinforced and reinforced samples versus freeze–thaw cycles.

134 M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

As seen, with increasing the number of freeze–thaw cycles the soilcohesion decreases, although there are somepoints inwhich the soil co-hesion increases. This suggests that voids between the soil particlesmayincrease due to forming ice lenses and thus the cohesion of soil reducesas this parameter is directly related to the distance between soil parti-cles. This phenomenon will be elaborated on the microfabric structuresection of the study.

These results are in agreement with the findings of Wang et al.(2007) who performed tests on unreinforced soil.

4.2.5. Friction angleThe internal friction angle represents the slipping and interlocking

features of one soil particle in relation to another. Triaxial test is the fun-damental test of obtaining the shear strength parameters of soil, cohe-sion and friction angle. According to ASTM D2850-03a, If the testspecimens are compacted when they are partially saturated, consolida-tion may occur when the confining pressure and deviatoric stress areapplied even drainage is not permitted. Therefore, if several partiallysaturated specimens of the samematerial are tested at different confin-ing stresses, they will not have the same undrained shear strength.Thus, the Mohr–Coulomb failure envelope for unconsolidated un-drained triaxial tests on partially saturated soils is usually curved anda friction angle will be obtained from the curves (ASTM, AmericanSociety for Testing Materials, 2003).

Table 3Final changes of the soil parameters after the ninth cycle.

A

3 = 90 kPa

A

3 = 60 kPa

A

3 = 30 kPa

Fiber

percentages

–0.16–0.27–0.430

–0.01–0.26–0.3170.5

–0.41–0.51–0.431

–0.28–0.4–0.3191.5

a The minus sign is for showing the reduction of a parameter.

Fig. 12 shows the changes of internal friction angle ratio of samplesduring freeze–thaw cycles. The ratio is defined as the friction angle ofa reinforced or unreinforced sample at a given cycle divided by that ofthe same sample which is not subjected to freeze–thaw cycles. The fric-tion angle values are denoted by φN and φ0, respectively.

As observed, although unreinforced samples experience a negligibleincrease in the friction angle during freeze–thaw cycles, the values offriction angle of reinforced samples remain relatively constant. The rea-son of this trend will be explained further in the next section.

Although enough tests have been performed to examine the repeat-ability of test results, the variability in some results of reinforced sam-ples is distinct. In order to find the optimum percentage of fibers, thefinal changes in parameters of the soil after the ninth cycle have beenpresented in Table 3. These changes can be expressed by

A ¼ 1−S9�S0

ð2Þ

B ¼ 1−C9�C0

ð3Þ

C ¼ 1−E9�E0

ð4Þ

where S9, C9, and E9 failure strength, cohesion and resilient modulus ofninth are cycle, respectively and S0, C0, and E0 are the initial values ofthese parameters before any cycle.

The highlighted numbers in the table are the least amount of chang-es in strength, cohesion and resilient modulus during the cycles.

It may be said that experimental errors and randomly distributed fi-bers may cause variability in results. Therefore, although some tenden-cies are observed, a number of them may only be due to the innervariability of results. This should be taken into account for interpretationof the final conclusion. The best reliable option for derivation of more ac-curate conclusion is to perform more tests for reproducibility in order toassess uncertainties in measured parameters. However, according to re-sults shown in Table 3, it seems that, by varying fiber percentage, some-times there is not a remarkable effect on varying parameters A, B, and C.However, in general, the amount of fibers about 0.5% may be more effec-tive than other percentages. Thus we cautiously introduce 0.5% fiber asthe optimum value for decreasing the effects of freeze–thaw cycles. Thisis because after applying 9 freeze–thaw cycles, parameter A of thesesamples is less than other values for two confining pressures.

4.2.6. Microfabric structureFreezing is a weathering process that occurs in thermodynamic con-

ditions just below 0 °C in which the soil moisture starts to freeze, ice

C

3 = 90 kPa

C

3 = 60 kPa

C

3 = 30 kPa

B

–0.62–0.76–0.49–0.73

–0.51–0.679–0.7–0.74

–0.84–0.681–0.78–0.46

–0.78–0.76–0.83–0.53

a-1 a-2 a-3

b-1 b-2 b-3

Fig. 13. SEM pictures of unreinforced samples (a) before and (b) after the cycles in three different magnifications.

135M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

crystals are formed and the soil is subjected to volumetric change. Thecrystal's volume increases up to 9 times and applies a considerablestress on soil aggregates, resulting in the change of soil characteristicin micro and macro scales (Konrad, 1989).

High soil moisture suction, developed in the zone of freezing, causesa very large increase in the effective stress in the unfrozen soil immedi-ately adjacent to this zone. It is known that consolidation occurs in frontof the 0 °C isotherm and in the unfrozen or partly frozen regions be-tween ice accumulations in freezing clay (Pusch, 1979).

Figs. 13–15 show the SEM pictures taken from unreinforced and re-inforced samples before and after the cycles in three different magnifi-cations (200, 800 and 5000 times bigger than the real dimensions).Fig. 13 (a-1 and b-1) depicts a very distinct increase in the amount ofbig pores of the pure soil samples after the cycles. The mineral skeletonadjacent to ice lenses and freezing front underwent appreciable com-pression. The compression itself caused by themovement of water dur-ing freezing develops a spatial network of channels which provideswater for the growing ice lenses. This network channel left afterthawing of ice crystals, increases the soil permeability. In unfrozen re-gions lying rather far from the freezing boundary, the porosity doesnot change much (Hohmann-Porebska, 2002).

It is worth mentioning that by developing the channels within thesoil skeleton, the distance between soil particles increases. This phe-nomenon is manifested in cohesion decrease after the cycles.

In addition, in picture b-1 of Fig. 13 coagulation of the soil particles isvisible. This phenomenon is the result of dehydration which occurs dueto themigration of freewater from the unfrozen part of soil to the freez-ing region during the freezing period.

In the second and third magnifications of Fig. 13 (a-2, b-2, a-3, b-3)the traces of ruptured primary particles and aggregates can be easily

distinguished by their broken contours and the abundance of fine-grained material along the boundaries of the elements which madetheir interaction increases. This phenomenon can be the reason of slightfriction angle increase during the cycles. Increase of friction angle canalso be related to the smoothness of the particles surface which is con-siderably distinct in SEM images.

In order to describe the microstructure changes after freeze–thawcycles accurately, the SEM images were transferred to AutoCAD soft-ware and some closed polygons were drawn around the distinct parti-cles which are in the surface of the pictures. Then, the area of eachparticle was calculated.

This procedure shows that after freeze–thaw cycles the size of parti-cles has been reduced about 12 times. These changes in dimension ofthe soil particles could be the result of compression caused by ice lensesto the mineral skeleton of the soil adjacent to frozen front area.

Moreover, Hohmann-Porebska (2002) illustrated that as the films ofboundwater around the particles become thinner andmore desiccated,the particles draw closer and their interaction increases, i.e., their struc-tural bonds become stronger. Therefore the friction angle augmentationafter the cycles which was stated in the experimental results of thisstudy can be explained bymicro-structural changes of the soil particles.

Figs. 14 and 15 show the micro-structural changes of the soil parti-cles in contactwith polypropylenefibers after the cycles in two differentmoisture contents (optimum and saturated). After the cycles, the sameprocess of augmentation of big pores in pure samples is equally recog-nizable in reinforced ones, especially in the first magnifications ofthese figures (c-1, d-1, e-1 and f-1). It is manifested in the cohesion de-crease between the fine grained soil particles and polypropylene fibers.The polypropylene fibers act as a tensile element between the soil par-ticles during the frost heave of freezing period. As the cohesion is the

c-1 c-2 c-3

d-1 d-2 d-3

Fig 14. SEM pictures of reinforced samples (c) before and (d) after the cycles in three different magnifications.

136 M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

sole connector between polypropylene fibers and soil particles, by de-creasing cohesion during the cycles, the tensile effect of the fibers willbe reduced. That is why increasing the amount of fibers in the soil sam-ples does not reduce the freeze–thaw influences. As the pictures c-2, 3,d-2, 3, e-2, 3 and f-2, 3 show, during freeze–thaw cycles the connectionof fibers and the soil particleswill decrease and so the tensile function offibers will be disrupted.

5. Conclusion

Laboratory tests and experimental investigations were conducted tomanifest the effect of polypropylene fiber existence in a fine grained soilduring freeze–thaw cycles. The results are as follows:

- Inclusion of 0.5%–1.5% polypropylene fibers results in increasing thefailure strength of fine grained soil to about 28%–65% and the fiber-reinforced soil exhibits more ductile behavior than the unreinforcedsoil.

- Soil cohesion increases about 67%–100% by adding fibers, but by in-creasing the fiber content this increment is reduced. The cohesiongrowth is a result of strength increase in the soil samples due tofiber addition and growth of plasticity. However, polypropylene fi-bers, because of their polymeric material, do not have significant co-hesion with the soil particles. So the cohesion reduction can bevisible by increasing the fiber content in the soil.

- By increasing the number of freeze–thaw cycles, the soil strength ofall reinforced and unreinforced samples decreases. However, thestrength decrease is more visible in unreinforced samples than in

reinforced ones. The stress–strain variation of thawed soil tends tovary from strain-hardening type to strain-softening type.

- The compression strength of unreinforced samples decreases by14%–43% due to the application of 9 freeze–thaw cycles while thisreduction is about 1%–32% for propylene-reinforced samples.

- By applying the freeze–thaw cycles, the resilientmodulus of both re-inforced and unreinforced samples decreases and using the polypro-pylenefibers in thefine grained soil does not have a significant effectin controlling the reduction of resilient modulus. The decreasedmagnitude of resilient modulus is more obvious in the first cyclesbutwhen the number of freeze–thaw cycles exceeds seven, the resil-ient modulus will reach a certain value and remains constant uponapplying further freeze–thaw cycles. This phenomenon shows anew equilibrium condition which becomes predominant on thesoil after 6–7 cycles as most of the changes occur at 1st to 7th cycles.

- Soil cohesion decreases with increasing the number of freeze–thawcycles. This suggests that voids between clay particles increase dueto forming ice lenses and thus the volume increases. The reductionin cohesion of soil is directly related to the distance increase be-tween soil particles after the cycles.

- Augment of big pores is recognizable after the cycles especially in thefirst magnifications of SEM pictures. It is manifested in the cohesiondecrease between the clay particles and polypropylene fibers.

- The values of friction angle of reinforced samples remain relativelyconstant after the cycles, although unreinforced samples experiencean increase in the friction angle. In SEM pictures the traces of rup-tured primary particles and aggregates can be easily distinguishedby their broken contours and the abundance of fine-grainedmaterial

e-1 e-2 e-3

f-1 f-2 f-3

Fig. 15. SEM pictures of saturated reinforced samples (c) before and (d) after the cycles in three different magnifications.

137M. Roustaei et al. / Cold Regions Science and Technology 120 (2015) 127–137

along the boundaries of the elements. This phenomenon can be thereason of friction angle increase during the cycles.

- SEM pictures of the soil samples before and after the cycles show co-agulation of the soil particles as a result of dehydrationwhich occursdue to the migration of free water from the unfrozen part of the soilto the freezing region during the freezing period.

- The polypropylene fibers act as a tensile element between soil parti-cles during the frost heave of freezing period. As the cohesion is thesole connector between polypropylene fibers and soil particles, bydecreasing cohesion during the cycles, the tensile effect of the fiberswill be reduced. That is why increasing the amount of fibers in thesoil samples does not reduce the freeze–thaw influences as theSEM pictures show that during freeze–thaw cycles the connectionof fibers and the soil particles will decrease and so the tensilefunction of fibers will be disrupted.

- Finally, according to the experimental results, the amount of fibersabout 0.5% may be more effective than other percentages and thisamount of fibers can be introduced as the optimum value for de-creasing the effects of freeze–thaw cycles.

References

Aldaood, A., Bouasker, M., Al-Mukhtar, M., 2014. Impact of freeze–thaw cycles on me-chanical behaviour of lime stabilized gypseous soils. Cold Reg. Sci. Technol. 99, 38–45.

ASTM (American Society for Testing Materials), 2003. Annual Book of ASTM Standardsvol. 03. American Society for Testing Materials.

Chamberlain, E.J., Iskander, I., Hunsiker, S.E., 1990. Effect of freeze–thaw on the perme-ability and macrostructure of soils. Proceedings of the International Symposium onFrozen Soil Impacts on Agriculture, Range, and Forest LandsSpecial Report 90–1.Cold Regions Research and Engineering Laboratory, Hanover, New Hampshire,U.S.A, pp. 145–155.

Cui, Z.D., He, P.P., Yang, W.H., 2014. Mechanical properties of a silty clay subjected tofreezing–thawing. Cold Reg. Sci. Technol. 98, 26–34.

Eigenbrod, K.D., 1996. Effects of cyclic freezing and thawing on volume changes and per-meabilities of soft fine-grained soils. Can. Geotech. J. 33, 529–537.

Ghazavi, M., Roustaie, M., 2010. The influence of freeze–thaw cycles on the unconfinedcompressive strength of fiber-reinforced clay. Cold Reg. Sci. Technol. 61, 125–131.

Ghazavi, M., Roustaie, M., 2013. Freeze–thaw performance of clayey soil reinforced withgeotextile layer. Cold Reg. Sci. Technol. 89, 22–29.

Graham, J., Au, V.C.S., 1985. Effects of freeze–thaw and softening on a natural clay at lowstresses. Can. Geotech. J. 22 (1), 69–78.

Güllü, H., Khudir, A., 2014. Effect of freeze–thaw cycles on unconfined compressivestrength of fine-grained soil treated with jute fiber, steel fiber and lime. Cold Reg.Sci. Technol. 106–107, 55–65.

Hazirbaba, K., Gullu, H., 2010. California Bearing Ratio improvement and freeze–thaw per-formance of fine-grained soils treated with geofiber and synthetic fluid. Cold Reg. Sci.Technol. 63, 50–60.

Hohmann-Porebska, M., 2002. Microfabric effects in frozen clays in relation to geotechni-cal parameters. Appl. Clay Sci. 21, 77–87.

Kalkan, E., 2009. Effects of silica fume on the geotechnical properties of fine-grained soilsexposed to freeze and thaw. Cold Reg. Sci. Technol. 58, 130–135.

Konrad, J.M., 1989. Physical processes during freeze–thaw cycles in clayey silts. Cold Reg.Sci. Technol. 16, 291–303.

Lee, W., Bohra, N.C., Altschaeffl, A.G., White, T.D., 1995. Resilient modulus of cohesive soilsand the effect of freeze–thaw. Can. Geotech. J. 32, 559–568.

Leroueil, S., Tardif, J., Roy, M., Konrad, J.-M., 1991. Effects of frost on themechanical behav-ior of Champlain Sea clays. Can. Geotech. J. 28 (5), 690–697.

Liu, J., Wang, T., Tian, Y., 2010. Experimental study of the dynamic properties of cement-and lime-modified clay soils subjected to freeze–thaw cycles. Cold Reg. Sci. Technol.61, 29–33.

Pusch, R., 1979. Unfrozen water as a function of clay microstructure. Eng. Geol. 13,157–162.

Qi, J., Vermeer, P.A., Cheng, G., 2006. A review of the influence of freeze–thaw cycles onsoil geotechnical properties. Permafr. Periglac. Process. 17, 245–252.

Shibi, T., Kamei, T., 2014. Effect of freeze–thaw cycles on the strength and physical prop-erties of cement-stabilised soil containing recycled bassanite and coal ash. Cold Reg.Sci. Technol. 106–107, 36–45.

Simonsen, E., Isacsson, U., 2001. Soil behavior during freezing and thawing using variableand constant confining pressure triaxial tests. Can. Geotech. J. 38, 863–875.

Wang, D., Ma,W., Niu, Y.H., Chang, X., Wen, Z., 2007. Effects of cyclic freezing and thawingon mechanical properties of Qinghai–Tibet clay. State Key Laboratory of Frozen SoilEngineering, Cold and Arid Regions Environmental and Engineering Research Insti-tute. Cold Reg. Sci. Technol. 48, 34–43.

Wang, T., Liu, Y., Yan, H., Xu, L., 2015. An experimental study on themechanical propertiesof silty soils under repeated freeze–thaw cycles. Cold Reg. Sci. Technol. 112, 51–65.

Yarbesi, N., Kalkan, E., Akbulut, S., 2007. Modification of the geotechnical properties, as in-fluenced by freeze–thaw, of granular soils with waste additives. Cold Reg. Sci.Technol. 48, 44–54.

Zaimoglu, A.S., 2010. Freezing–thawing behavior of fine-grained soils reinforced withpolypropylene fibers. Cold Reg. Sci. Technol. 60, 63–65.